Article Text

Download PDFPDF

Original research
Clinical and subclinical microemboli following neuroangiography in children
  1. Ibrahim Alghamdi,
  2. Adam A Dmytriw,
  3. Afsaneh Amirabadi,
  4. Samantha Lebarron,
  5. Vanessa Rea,
  6. Carmen Parra-Fariñas,
  7. Prakash Muthusami
  1. Divisions of Neuroradiology & Neurointervention, Hospital for Sick Children, University of Toronto, Toronto, ON, Canada
  1. Correspondence to Dr Prakash Muthusami, Divisions of Neuroradiology & Neurointervention, Hospital for Sick Children, University of Toronto, Toronto, ON, Canada; prakash.muthusami{at}


Background To assess the frequency, imaging appearances, and risk factors of brain microemboli following pediatric neuroangiography, as assessed by early diffusion-weighted MRI imaging (DWI).

Methods This single-center, retrospective analysis investigated early DWI post-pediatric neuroangiography. Patients aged 0–18 years who had diagnostic neuroangiography and DWI within a week postprocedure were included. Data on clinical and procedural parameters and MRI findings were recorded. Univariate and multivariate analyses were performed on the following risk factors: age, weight, vasculopathy, antiplatelet drug use, access type, intraprocedural heparin, procedure duration, neck arteries catheterized, and angiographic runs. A p-value<0.05 indicated statistical significance.

Results Eighty-two children were included (40.2% female), mean age 10.1±4.5 years (range: 7 months–17 years). There were no intraprocedural thromboembolic complications recognized. DWI positivity was seen following 3.6% (3/82) procedures: two with transient symptoms, and one instance of silent microemboli. There were no territorial infarcts or clinical stroke. Children with underlying vasculopathy had a higher risk of microemboli from angiography than children without vasculopathy (OR 31.6, p=0.02), and the OR of microemboli following transradial angiography was 79.1 (p=0.005) as compared with transfemoral angiography. Univariate and multivariate analysis showed a significant association between microemboli and number of angiographic runs (p=0.004). Follow-up MRI in all three patients showed no residual abnormal signal.

Conclusions Cerebral microemboli are unusual following uncomplicated neuroangiography in children. However, in the presence of underlying vasculopathy and with transradial technique, the incidence approaches that reported in the adult literature. An increased association with the number of angiographic runs is an important and controllable factor.

  • stroke
  • angiography
  • catheter
  • MRI
  • complication

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Catheter-directed angiography is a crucial diagnostic tool for brain vascular pathologies in children and remains the gold standard. Although it has been proven safe, certain procedural risks remain, including access-related complications and the occurrence of brain microemboli, which are often subclinical and more frequently reported in adults undergoing neuroendovascular procedures.


  • This study uncovers that the occurrence of diffusion-restricting brain lesions due to microemboli following pediatric diagnostic angiography is relatively rare but becomes significantly more common in the presence of underlying vasculopathy. It also highlights that transradial angiography, in the presence of vasculopathy, further increases the risk of brain microemboli.


  • This study could prompt further research into assessing the role of procedural techniques, antiplatelet therapies, and procedural anticoagulation in pediatric patients with vasculopathy. The findings offer important information for deciding thresholds for performing pediatric cerebral angiography, for determining practice standards, and for advising families regarding procedural risks in an established pediatric practice. Additionally, this study could influence policy by promoting the development of guidelines on performing angiography for brain vascular pathologies in pediatric patients.


Catheter-directed cerebral angiography remains the gold standard to evaluate vascular pathology in the brain. With improvement in techniques, devices, and dedicated teams, this procedure can be performed with high accuracy and safety in high-volume pediatric centers with multidisciplinary support.1 2 Risks associated with intracranial endovascular procedures are well described in the literature, including vessel damage, dissection, bleeding, and neurological complications. Another known but less understood phenomenon is that of “microemboli”, which block tiny brain arteries and are described to occur following 11.15% to 100% of catheter cerebral angiographic procedures in the adult literature.3–5 This phenomenon and its frequency have been variably suggested to be related to underlying vascular disease, operator experience and technique, viscosity of contrast media injected, and the use of anticoagulation. In adults, the frequency and number of microemboli are more in prolonged interventions than during diagnostic angiography.6 7 Microemboli following pediatric cerebral angiography have not been studied, although it is believed that these should occur in lower frequency in children, given the lesser atherosclerotic burden and relatively straighter arteries. Children, however, have other risk factors that could result in microemboli, including smaller vessel to catheter ratios, distinct arteriopathies that affect cervical and intracranial vessels, neck vessel tortuosity, increased propensity for procedure-related vasospasm, hypotension from general anesthesia, and variable operator experience for pediatric procedures.

Given the increasing role of, and reliance on, catheter-directed angiography for pediatric neurovascular disease, it is essential to understand the frequency, imaging appearances, and risk factors of cerebral microemboli following neuroangiographic procedures in children.

We undertook a retrospective study looking at postprocedural cerebral microemboli in children who underwent MRI within a week of catheter-directed cerebral angiography at our quaternary pediatric institute.


Patients were enrolled in this retrospective study after institutional review board approval (#1000074427) with a waiver for individual consent. Patients under 18 years of age who had undergone diagnostic neuroangiography at our pediatric hospital over a 3-year period from July 1, 2018 to June 30, 2021, and who also received an MRI with DWI within 7 days following the procedure, were included in the study. Neuroangiography performed as part of interventional procedures was excluded. Procedures performed within 7 days of symptom onset were considered ‘acute’ and others categorized ‘elective’.

All procedures were performed under general anesthesia by a single neurointerventional operator with or without a fellow assistant in a biplane neuroangiography suite (Artis Q BA Twin; Siemens). Transfemoral neuroangiography was performed through a 10 cm long 4Fr vascular access sheath (Glidesheath; Terumo, Tokyo, Japan), using a 4Fr Berenstein tip diagnostic catheter (Merit Medical) and 0.035 inch hydrophilic guidewire (Glidewire; Terumo). Systemic heparinization (50 IU/kg, up to a maximum of 3000 IU) was used, except in acute hemorrhagic or postoperative cases. Transradial neuroangiography was performed through a 25 cm 5Fr Glidesheath Slender (Terumo), using a 5Fr Sim-1 Glide catheter (Terumo). Digital subtraction angiography (DSA) was performed using an iodinated contrast agent (Iohexol, Omnipaque 300 mg I/mL; GE Healthcare, Piscataway, NJ, USA) with hand injections. Catheters were maintained on continuous perfusion of heparinized saline (2 IU/mL). Flushes were drawn from open, plastic, single-use bowls on a separate surface. Wet Telfa gauze was used to wipe wires. Postprocedure hemostasis was achieved with manual compression for transfemoral procedures and an inflatable wrist band (TR Band; Terumo) for transradial procedures. Discharge following angiography was done after 6 hours of awake observation.

MRI was performed using a pediatric 8- or 16-channel head coil on a 1.5 T magnet (Signa Excite HD; GE Medical Systems, Milwaukee, WI, USA) or 3 T magnet (Magnetom Skyra; Siemens, Erlangen, Germany; or Achieva; Philips Healthcare, Best, Netherlands) with a 32-channel head coil. The protocol included multiplanar T1- and T2-weighted sequences followed by time of flight-magnetic resonance angiography (TOF-MRA) with institutional parameters described previously.8 DWI was performed in the axial plane with the following parameters: TR 6480 ms, TE 64 ms, flip angle 180°, bandwidth 919 Hz, FOV 220 mm, with 4 mm slice thickness and b-values of 0 and 1000 s/mm2.

Patient demographics and clinical data were retrieved from the electronic medical record and imaging data from the picture archiving and communication system (PACS). Clinical parameters assessed (preprocedure) included age, weight, sex, indication for catheter angiography, symptoms at presentation, any previous strokes, prescription medications (antiplatelets, anticoagulants, immunosuppressants), and the presence and type of underlying vasculopathy. The diagnosis of vasculopathy was a clinico-radiologic one, and included a wide spectrum of inflammatory, genetic, or idiopathic conditions that cause vessel narrowing or occlusions by affecting the vessel wall rather than primarily due to luminal occlusion by clot or emboli. This includes conditions like focal cerebral arteriopathy, sickle cell disease, moyamoya disease, Menkes disease, PHACE syndrome, ACTA2 mutation, collagen vascular diseases like Ehlers–Danlos type IV, fibromuscular dysplasia, and post-infectious arteriopathy.

Procedural parameters assessed included access location (femoral vs radial), duration of the procedure, fluoroscopy time, number of angiographic runs in head and neck vessels, number of neck vessels catheterized, heparin use, and procedure-related complications. Clinical parameters assessed (postprocedure) included new symptoms after the procedure and before MRI lesions. MRI images were analyzed by a radiologist blinded to clinical features and to angiography procedural findings or outcome. MRI parameters assessed included indication for MRI, interval between procedure and MRI, presence, number, and location of new DWI lesions, follow-up MRI, the interval between initial and follow-up MRI, and persistence of abnormal signal. Microemboli on DWI were defined as high focal diffusion trace intensity with corresponding apparent diffusion coefficient (ADC) hypointensity, measuring <10 mm in maximum diameter, either cortical, subcortical, or lacunar. DWI lesions on postprocedural MRI that were already present on preprocedural MRI were excluded. Lesions larger than 10 mm and in a classical arterial territory were classified as infarcts.

Estimating a summary incidence of 25% from the reported adult literature on DWI microemboli, a required sample size of 80 was calculated to identify half this incidence in children, to power the study to 80% with a type I error (α) of 0.05. Descriptive statistics were used for clinical and imaging parameters. The association between DWI lesion presence with clinical (age, weight, vasculopathy, prescription medications) and procedural (use of intraprocedural heparin, procedure time, and the number of angiographic runs) parameters was performed using univariate and multivariate logistic regression. Chi-square test for proportions was used for comparing lesion frequency against the presence/absence of vasculopathy. A p-value<0.05 was considered statistically significant.


Of a total of 322 pediatric neuroangiographic procedures performed in the study period, 82 patients (25.4%) met inclusion criteria. Among these, 33 were female (40.2%). The mean age of the patients was 10.1±4.5 years (range: 7 months–17 years). Underlying vasculopathy was present in 17 patients (20.7%), most commonly moyamoya disease; 3/17 of these patients were on antiplatelet therapy with aspirin alone or with clopidogrel. No patient was on steroids or immune suppression at the time of the procedure. Demographics are summarized in table 1.

Table 1

Patient demographics in our cohort (N=82).

Angiography was performed within 7 days of symptom onset in 42/82 (51.2%) patients and electively in the others. Transfemoral procedures (73/82) accounted for 89% of the cases, while radial access was used in the remaining 11%. Mean fluoroscopy time for diagnostic neuroangiography was 6.7±13.0 minutes. Intraprocedural heparin was used in 41/82 (50%) of procedures, with a mean heparin dose of 2080 IU (range: 300–3000 IU). All the patients with vasculopathy received heparin for angiography.

There were no intraprocedural thromboembolic complications recognized. None of the cases had intraprocedural vasospasm requiring intra-arterial vasodilator therapy. There was no postprocedural access site hemorrhage or limb ischemia requiring ultrasound, change of care plan, or admission. Postprocedural neurological symptoms occurred in two patients, both with underlying moyamoya arteriopathy, who developed transient hand weakness and blurry vision, respectively, following uneventful elective diagnostic angiography. MRI was performed for both patients, both positive for multiple DWI microemboli (figure 1). Indications for MRI in the other 80 children are shown in table 1. The interval between the procedures and MRI was 2.6±2.1 days (range: 1–7 days). One of these 80 patients had a diagnosis of focal cerebral arteriopathy and recent previous stroke, and had diffusion positivity for new microemboli, although asymptomatic.

Figure 1

A paedatric patient with moyamoya disease who developed transient blurry vision following diagnostic angiography. (A) Digital subtraction angiogram (DSA) of the right internal carotid artery demonstrating severe stenosis of the M1-middle cerebral artery. (B) Left internal carotid artery injection showing no disease on this side. (C) Axial diffusion-weighted imaging (DWI) image from MRI on the same patient post-diagnostic angiography demonstrating multiple scattered foci of restricted diffusion involving left parietal and frontal lobes. Note that right-sided occipital lesions are not visible in this section. (D) Follow-up fluid-attenuated inversion recovery (FLAIR) image 1 month later showing interval evolution of the lesions.

In total, DWI microemboli were seen following 3/82 (3.6%) diagnostic cerebral angiograms. All three DWI-positive studies were in children having an underlying vasculopathy (3/17=17.6%) as compared with 0/65 in non-vasculopathic children (p<0.0001). All these three children with DWI positivity had received transradial angiography under systemic heparinization, and one was on long-term antiplatelet treatment with clopidogrel. This resulted in an OR of 31.6 (p=0.02) for DWI microemboli following cerebral angiography in the presence of an underlying vasculopathy, increasing to an OR of 79.1 (p=0.005) when angiography was performed via the transradial route. None of the transfemoral procedures had DWI positivity on MRI.

In the univariable logistic regression model for risk factors (age, weight, known vasculopathy status, preprocedural medications, use of intraprocedural heparin, procedure time, fluoroscopy time, and number of DSA runs), only the number of DSA runs (OR 1.15, 95% CI 1.05 to 1.26, p=0.002) were significantly associated with the presence of DWI microemboli. On multivariate analysis, this remained significantly associated (OR 1.2, 95% CI 1.05 to 1.31, p=0.004) with microemboli on postprocedural MRI. These results are summarized in table 2.

Table 2

Univariate and multivariate analysis of association of clinical and imaging parameters with presence of diffusion-weighted imaging lesions on magnetic resonance imaging

All three patients with DWI positivity had a follow-up MRI at 21.3±15.0 days interval, which showed no residual abnormal fluid-attenuated inversion recovery (FLAIR) signal, encephalomalacia, and no new lesions (table 3). No new neurological symptoms were present.

Table 3

Details of the three children with diffusion-weighted imaging-positive microemboli following catheter-directed cerebral angiograpghy


Catheter-directed angiography is the gold standard test for brain vascular pathology in children despite the excellent and improving quality of current non-invasive cross-sectional modalities.9 Furthermore, the continuous improvement of angiographic techniques such as 3D rotational angiography and 4D-CT angiography add more value in diagnosing subtle pathology and understanding hemodynamics of the intracranial circulation.10 11 However, despite these advantages, there remain inherent and often unpredictable procedural risks thus far not precisely documented in the pediatric literature.

Cerebral angiography in children is proven to be safe with a low rate of permanent or major complications when performed in large centers with pediatric expertise.1 Access-related complications, including groin hematoma and vascular dissections, are among the most common associated risks with the invasive nature of the procedure.12 13 Brain microemboli, detected by diffusion-weighted MRI, are well described in the angiographic literature following neuroendovascular procedures in adults, and are most often subclinical (silent).14 In fact, microembolic events are reportedly far more frequent than the neurological complication rate. Bendszus et al7 assessed the neurological complications and the frequency of silent microembolism after cerebral angiography in 100 procedures in adult patients. Positive DWI microembolic lesions were found in approximately a quarter of these patients, not associated with clinical symptoms (ie, were ‘silent’). Similarly, in a study by Kato et al8 on 50 patients who underwent cerebral angiography, 13/50 patients (26%) had new DWI embolic lesions after the procedure, none of them symptomatic. An even higher incidence of silent thromboembolic events and neurological events was found to be associated with therapeutic neuroendovascular procedures like cerebral aneurysm coiling and carotid arterial stenting, as high as 40% to 60% in some studies.15 16

The sensitivity and usefulness of DWI in detecting early cerebral infarcts have been shown in several studies.17 18 DWI and ADC maps are well established as the gold standard for detecting brain ischemia, with 94% to 100% sensitivity within 6 hours of symptom onset. Although ADC maps show “pseudonormalization” after 7–10 days, signal changes on b-1000 DWI images can persist for up to 2–4 weeks19 depending on the duration, severity, and extent of the initial insult.20 21 These signal abnormalities detected on DWI are believed to be related to microthrombi, air bubbles, fibers from drapes or cotton gauze, insoluble dried contrast media residue, and other foreign body materials shed by catheters, guidewires, syringes, stopcocks, and other devices during the procedure.22 This is corroborated by the increased incidence in patients with vascular risk factors for atherosclerosis, with longer procedures and with more catheter maneuvers or exchanges. This has also been suggested based on transcranial Doppler studies.4 Although uncommon, transient and permanent neurological complications have been reported.13 23 Cerebral microinfarcts, including those that are clinically silent and detected24 on DWI25 and FLAIR,26 have also been implicated as an important underlying risk factor for dementia and cognitive impairment27 28 in adult populations. In addition, several studies have reported a short- and long-term cognitive decline related to DWI lesion burden following neurointerventions.29 30

Since DWI microemboli are commonly silent, their non-occurrence in children cannot be assumed. Overall, DWI lesions occurred in 3.6% of our retrospective cohort, which is considerably less than the reported rate in adults. The incidence, however, was more in the presence of vasculopathy; 20.7% of the children in our study had an underlying vasculopathy and had a 17.6% incidence of (clinical and silent) microemboli as compared with 0% in children without vasculopathy (OR 31.6). None of these DWI lesions showed residual signal abnormality on follow-up imaging to suggest development of lacunar or subcortical infarcts. This is in keeping with the literature, which suggests that long-term outcomes of DWI microemboli depend on the size of the initial lesion.31 32 Zhou et al showed that following carotid interventions in adults, residual MRI abnormalities are seen only when initial DWI lesions are larger than 60 mm2.33 This incidence of DWI positivity in vasculopathic children is as high as that described in the adult literature following diagnostic cerebral angiography,3 4 7 suggesting that the etiology of these lesions is likely intrinsic to vascular disease rather than inherent to technique. In addition, the trend we found with the number of angiographic runs performed suggests an interaction of vascular pathology with technique. A prospective randomized control trial showed a significant reduction in microemboli detection by DWI with air filters and systemic heparin use compared with the control group.34 Nonetheless, patients with underlying vascular risk profiles were associated with higher chances of having microemboli.7 35 One of the three children with DWI microemboli in our study was on clopidogrel monotherapy, and the other two received intraprocedural heparin. However, the numbers were inadequate for statistical analysis of the use of anticoagulation or antiplatelet agents, and these would require specifically designed studies addressing this issue.

Interestingly, all three children with DWI microemboli had also received transradial neuroangiography, whereas none of the children with transfemoral neuroangiography had DWI microemboli (with an OR of 79.1). This increased risk of cerebral microemboli from transradial angiography has also been described in the coronary angiography literature,36 and is believed to be related to dislodgment of atherosclerotic plaques in that population. Although this is not a risk factor in children, the increased catheter manipulation with a reverse curve catheter required during transradial angiography could be an inciting factor for endothelial microtrauma and microthrombi entering the intracranial circulation in children with vasculopathy. Given the increasing popularity of transradial neuroangiography especially for patients on blood thinners, this suggestion for an increased propensity to microemboli needs to be further assessed with larger samples.

There are some limitations present in our study. The retrospective nature of the study introduces an expected selection bias, since only a quarter of all procedures performed during this period were included. This would skew the results towards symptomatic patients who received MRI. This could also underrepresent silent microthrombi in an unselected cohort of all children receiving angiography, particularly those with underlying vasculopathy, and also younger children and infants who would not report symptoms and be more difficult to assess. On the contrary, it is likely that the study selection is biased towards patients who received angiography for stroke work-up or preoperative work-up, since these groups are more likely to receive MRI for stroke follow-up and postoperative assessment, respectively. This group would also be more likely to have new microemboli from the underlying disease rather than from interventions. Second, the study was not designed to adequately examine some risk factors, including the routine use of anticoagulation during pediatric neuroangiography, use of antiplatelet agents in vasculopathic patients, and interoperator differences. These factors require further investigation with appropriately designed and powered studies. The study sample was heterogeneous, with several vascular diseases included for angiography. From the results of the study it would have been important to look at a larger population of children with vasculopathy undergoing angiography, and the risk factors therein. However, we believe that this study retains its clinical relevance as it does represent the actual spread of cases one could expect in a pediatric neuroangiographic practice and serves to identify items for further exploration. Lastly, although all DWI were performed using the same techniques, the variable postprocedural timing of MRI might introduce variation in the detection of early DWI lesions. However, none of the MRI studies were late enough to have b-1000 image pseudonormalization.

In conclusion, we found that diffusion-restricting brain lesions due to microemboli are an unexpected finding following uncomplicated diagnostic angiography in children. However, in the presence of vasculopathy, their incidence approaches that reported in adults. As a corollary, when diffusion-restricting brain lesions due to microemboli occur following uncomplicated angiography, this should trigger an investigation for underlying vasculopathy. Additionally, in the presence of vasculopathy, transradial angiography considerably increases the risk of brain microemboli. The above observations are important information for deciding thresholds to perform pediatric cerebral angiography, for determining practice standards, and for advising families regarding procedural risks in an established pediatric practice. Further studies are required within this group to assess the role of specific procedural techniques, antiplatelet therapies, and procedural anticoagulation.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

Hospital for Sick Children IRB #13-03-2020.



  • Twitter @ibrahim_Gh, @AdamDmytriw

  • Contributors Conception or design of the work: IA, AAD, AA, SL. Data acquisition and analysis: IA, AAD, SL, AA. Interpretation of data: IA, AAD, VR, PM, CP-F, AA. Drafting the work: IA, AAD, AA, CP-F, PM. Revising the work for valuable intellectual content: IA, SL, VR, AAD. Approval of the final version: all authors. PM is the guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.